Council of Europe opens AI convention for signature

On 5 September 2024, the Council of Europe’s Framework Convention on Artificial Intelligence, Human Rights, Democracy, and the Rule of Law will be officially opened for signature during an informal Conference of the Ministers of Justice of the Council of Europe in Vilnius, Lithuania. The Convention, adopted on 17 May 2024, during the Council of Europe’s Committee of Ministers’ annual meeting, provides a global legal framework for regulating AI systems, with a focus on ensuring these technologies align with human rights, democratic integrity, and the rule of law.

The Convention is set to guide countries on how to manage the development and implementation of AI within the public sector, ensuring that AI technologies respect essential democratic values. Importantly the Convention also covers private-sector entities when they operate on behalf of public authorities, such as under government contracts, further broadening its scope of application.

Key Principles for Adoption

The Framework Convention establishes seven core principles for implementing AI governance, which signatories need to adapt to their domestic legal frameworks. These principles include:

• Human dignity and individual autonomy: Ensuring AI does not undermine human dignity or reduce individuals to mere data points.
• Transparency and oversight: Maintaining transparency across the AI lifecycle, ensuring users can understand AI-generated content and decisions.
• Accountability: Clear responsibility for AI’s impact on human rights, including legal mechanisms to hold actors accountable.
• Equality and non-discrimination: Addressing and mitigating bias throughout the AI system’s lifecycle.
• Privacy and data protection: Safeguarding individuals’ privacy and controlling access to personal data.
• Reliability: Ensuring AI systems are robust, secure, and trustworthy in their outputs.
• Safe innovation: Promoting innovation that respects democratic principles and human rights, with support for regulatory sandboxes and clear guidance.

The Convention allows for national security exemptions, i.e. the Parties are allowed to not implement the treaty for activities protecting national security, provided that these comply with international law.

Global impact

The Convention has been hailed (by some) as a major achievement in global AI governance. Its goal is to harmonise the regulation of AI systems across borders and create a framework that can be adopted by both European and non-European countries. The Treaty has already garnered significant international interest, with nations outside Europe, including the US, Canada, Japan, and Australia, participating in negotiations alongside private sector and civil society observers.

Human rights and democracy

A key aspect of the Convention is its focus on safeguarding human rights and democracy against the potentially harmful effects of AI technologies. It mandates that AI systems should always maintain human autonomy, preventing the dehumanisation of individuals and ensuring their ability to exercise control over AI’s influence in their lives. The document also emphasises the importance of transparency, urging the use of content labelling, watermarking, and other methods to ensure that users can clearly distinguish between human-generated and AI-generated content.

Criticisms

Despite its historic significance, some view the treaty as falling short of its intended impact. Concerns have been raised regarding the treaty’s effectiveness, with suggestions that it primarily reaffirms existing practices rather than introducing substantive regulatory measures.

The EU data watchdog expressed concerns about potential compromises in human rights standards due to pressure from foreign business interests. The European Data Protection Supervisor (EDPS) believes that the proposal does not go far enough in addressing the risks and challenges posed by AI. The text has been significantly weakened from its original version during negotiations at the CoE’s ad hoc committee in charge of the convention. The EDPS described it as a ‘missed opportunity to lay down a strong and effective legal framework’ for protecting human rights in AI development.

Why does it matter?

The opening of the Framework Convention for signature in Vilnius marks an important moment for AI governance. While the Convention sets out an ambitious agenda, its success depends on the willingness of signatory countries to implement its principles through their own legislative and regulatory systems. If widely adopted, this could serve as a global benchmark for AI governance, potentially inspiring similar frameworks in other regions.